The adenovirus E1A proteins are capable of exerting a wide range of functions, including activation and repression of many promoters. This is apparently related to E1A's ability to interact with a variety of cellular factors that play key roles in regulating gene expression. Transactivation by Ad5 E1A: The 46 amino acid transactivating domain of the 289R E1A protein of Ad5 contains a Cys4 zinc finger structure which is required for transactivation. E1A transdominant mutants, generated by mutating residues which are directly carboxyl to the zinc finger, are thought to sequester a cellular factor required by wild type E1A. From genetic analysis, one model predicts that the zinc finger and the adjacent carboxyl regions bind to different transcription factors. Indeed, TFIID and ATF are now known to bind E1A's transactivating domain. Our aim is (i) to use our many E1A substitution mutants in binding studies to test the prediction that TFIID associates with the finger region, While ATF binds to the adjacent carboxyl region and (ii) to gain a greater understanding of E1A and its transactivating domain structure through analysis of E1A crystals, and possibly E1A-TFIID co-crystals. Repression of MHC class I by Adl2 ElA: Adl2 (but not Ad5) transformed cells are, tumorigenic and exhibit diminished class I expression, which may help them escape immunosurveillance by CTLs. Adl2 E1A mediates the block in class I expression at the level of transcription and the R2 element of the class enhancer is a target. The strong binding activity to the R2 element (Complex A) correlates with the existence of an apparent trans-negative factor and with the ability of R2 to singly repress the positive enhancer element, R1. A hormone receptor protein (probably, RXRbeta) is implicated to be part of Complex A because the binding activity is lost by mutating the hormone receptor recognition sequence in R2 and because retinoic acid (RA) overrides R2 repression. In this study we aim (i) to determine the nature of Complex A by DNA footprint and methylation interference assays (with and without RA); to assess the presence of RXRbeta and associated proteins using antibodies in band-shift assays; and to identify unknown proteins that bind directly to R2 through screening an Adl2 cDNA expression library (ii) to determine if R2 can repress transcription independently of Rl and (iii) to fine map a domain of Adl2 E1A that may be required for repressing class I transcription and to test this domain in tumorigenesis.